When interactive advantage reverses: AI, authenticity erosion and dynamic multi-channel performance

Purpose Brands today face a striking managerial anomaly: artificial intelligence (AI)-driven content volume has surged by over 35% year-on-year, yet engagement rates have stagnated or declined. This study asks why early interactive performance gains so reliably reverse and identifies two novel findings that resolve this puzzle. First, authenticity erosion does not merely accompany relationship quality decline; it precedes and amplifies it, serving as the earliest detectable warning of relational reversal. Second, a higher AI-content share compresses the temporal window between initial performance gains and their reversal, accelerating the curvature of the interactive advantage trajectory. Design/methodology/approach A two-study, multi-method design integrates a natural language processing (NLP)-driven event-study difference-in-differences (Study 1) to trace temporal escalation effects on relational tone, authenticity and performance, with dynamic system generalized method of moments (GMM) panel modelling (Study 2) to uncover lagged mechanisms involving interactive intensity, AI-content share, relationship quality and authenticity. Findings Interactive escalation follows a non-monotonic inverted-U trajectory: initial gains give way to relational reversal as authenticity erodes, parasocial bonds weaken and relationship quality declines. Authenticity erosion is the leading signal of this reversal, preceding broader relational decline. Higher AI-content share compresses the gain-to-decline window, steepening and advancing the curvature through a negative interaction between AI share, interactive intensity and perceived authenticity. Originality/value Unlike prior static models assuming linear/diminishing returns or classical wear-out, this research formalises “reversible advantage” as a path-dependent process, with AI accelerating curvature. It bridges theory-practice gaps by offering diagnostic pacing heuristics, advancing interactive marketing with empirical evidence on temporal advantage cycles in AI ecosystems.